A Smaller Workforce Requires Better Work Design, Not Just New Headcount Plans

A smaller workforce producing more output creates a planning problem that headcount targets will not solve. For a CHRO or chief people officer, the question is not simply how many roles AI may remove or augment by 2035. It is whether changed human capability will alter the jobs, decisions and operating rules around those roles. Workforce planning will lose value when it stays focused on projections and leaves work design unchanged.

A recent analysis of a new US Bureau of Labor Statistics occupational forecast describes a 2035 workforce that is smaller, more productive and in high demand. That is the external fact. It does not establish how productivity will change in a particular sector, or which occupations will expand, contract or be redesigned. It does create a planning pressure: organizations may need more output from fewer people while competing for capabilities that remain scarce.

The usual response will be to revise workforce numbers, job descriptions and hiring plans. Those changes may be necessary, but they will not resolve every constraint. An employee might use AI to prepare a sound first analysis in an hour rather than a day, only for the gain to disappear in the same approval route, manager queue or separation of duties. A role built mainly around gathering information may also stop giving people enough practice in interpretation and judgment. Capacity can rise while the organization remains built for yesterday.

I would start with the work rather than the job title. For a priority workflow, map the outcome, the decisions that protect quality, the tasks AI could assist with if adopted, and the points where a person must still verify, interpret or decide. Compare that map with the current role profile, manager span, approval path and time to proficiency. The executive owner may be the business leader, supported by HR, workforce planning and the learning function. Before funding another course or opening another requisition, determine whether the constraint sits in skill, process, authority or governance.

A lower staffing requirement is not automatically a productivity gain. Experienced employees may absorb more review work, managers may become a bottleneck, and early career staff may lose opportunities to practise judgment. In an AI enabled workflow, the useful test is whether redesigned work improves throughput while preserving the human capability needed for accountable decisions. Some review and approval controls should remain even when production becomes faster.

The forecast cannot tell us whether that redesign will happen. Productivity effects will vary by occupation and sector, and there is limited evidence for precise 2035 claims about individual roles. The source supports a planning hypothesis, not a market wide conclusion. Leaders should run small, measurable tests in important workflows before changing the staffing model. Record cycle time, rework, manager effort, quality decisions and time to proficiency.

Improvement across those measures would give the organization evidence to adjust roles and investment. If output rises while review load, errors or lost development opportunities rise with it, the workforce plan is masking an operating design problem. Every AI related headcount assumption should therefore rest on a tested change in work, authority and accountability. Projected productivity is not capacity until the new way of working delivers it.